# When Does RL Help Medical VLMs? Disentangling Vision, SFT, and RL Gains

arXiv: https://arxiv.org/abs/2603.01301 · HTML: https://mlsanity.com/paper/2603.01301/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Clinical and biomedical applications, one of 17 directions in LLM reasoning.
- Builds on 28 of its 33 references that are in the map.
- First posted on arXiv 1 Mar 2026.

## Placed in

- LLM reasoning / [Clinical and biomedical applications](https://mlsanity.com/lr/clinical-and-biomedical-applications/)
- Multimodal LLMs / vision-language models / [Medical vision-language models](https://mlsanity.com/mllm/medical-vision-language-models/)

## Builds on (most cited first, 28 of 28)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI](https://mlsanity.com/paper/2311.16502/) (2023)
- [Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?](https://mlsanity.com/paper/2504.13837/) (2025)
- [MedMNIST v2 -- A large-scale lightweight benchmark for 2D and 3D biomedical image classification](https://mlsanity.com/paper/2110.14795/) (2021)
- [PathVQA: 30000+ Questions for Medical Visual Question Answering](https://mlsanity.com/paper/2003.10286/) (2020)
- [SLAKE: A Semantically-Labeled Knowledge-Enhanced Dataset for Medical Visual Question Answering](https://mlsanity.com/paper/2102.09542/) (2021)
- [PMC-VQA: Visual Instruction Tuning for Medical Visual Question Answering](https://mlsanity.com/paper/2305.10415/) (2023)
- [Lingshu: A Generalist Foundation Model for Unified Multimodal Medical Understanding and Reasoning](https://mlsanity.com/paper/2506.07044/) (2025)
- [MedXpertQA: Benchmarking Expert-Level Medical Reasoning and Understanding](https://mlsanity.com/paper/2501.18362/) (2025)
- [MedVLM-R1: Incentivizing Medical Reasoning Capability of Vision-Language Models (VLMs) via Reinforcement Learning](https://mlsanity.com/paper/2502.19634/) (2025)
- [Reasoning with Sampling: Your Base Model is Smarter Than You Think](https://mlsanity.com/paper/2510.14901/) (2025)
- [On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models](https://mlsanity.com/paper/2512.07783/) (2025)
- [GRPO-CARE: Consistency-Aware Reinforcement Learning for Multimodal Reasoning](https://mlsanity.com/paper/2506.16141/) (2025)
- [Med-RLVR: Emerging Medical Reasoning from a 3B base model via reinforcement Learning](https://mlsanity.com/paper/2502.19655/) (2025)
- [Patho-R1: A Multimodal Reinforcement Learning-Based Pathology Expert Reasoner](https://mlsanity.com/paper/2505.11404/) (2025)
- [MMedAgent-RL: Optimizing Multi-Agent Collaboration for Multimodal Medical Reasoning](https://mlsanity.com/paper/2506.00555/) (2025)
- [QoQ-Med: Building Multimodal Clinical Foundation Models with Domain-Aware GRPO Training](https://mlsanity.com/paper/2506.00711/) (2025)
- [Beyond Distillation: Pushing the Limits of Medical LLM Reasoning with Minimalist Rule-Based RL](https://mlsanity.com/paper/2505.17952/) (2025)
- [ChestX-Reasoner: Advancing Radiology Foundation Models with Reasoning through Step-by-Step Verification](https://mlsanity.com/paper/2504.20930/) (2025)
- [MedVLThinker: Simple Baselines for Multimodal Medical Reasoning](https://mlsanity.com/paper/2508.02669/) (2025)
- [MedGround-R1: Advancing Medical Image Grounding via Spatial-Semantic Rewarded Group Relative Policy Optimization](https://mlsanity.com/paper/2507.02994/) (2025)
- [Medical Reasoning in the Era of LLMs: A Systematic Review of Enhancement Techniques and Applications](https://mlsanity.com/paper/2508.00669/) (2025)
- [OctoMed: Data Recipes for State-of-the-Art Multimodal Medical Reasoning](https://mlsanity.com/paper/2511.23269/) (2025)
- [Advancing Medical Representation Learning Through High-Quality Data](https://mlsanity.com/paper/2503.14377/) (2025)
- [Reasoning Visual Language Model for Chest X-Ray Analysis](https://mlsanity.com/paper/2510.23968/) (2025)
- [Medical Image Classification with KAN-Integrated Transformers and Dilated Neighborhood Attention](https://mlsanity.com/paper/2502.13693/) (2025)
- [PuzzleCraft: Exploration-Aware Curriculum Learning for Puzzle-Based RLVR in VLMs](https://mlsanity.com/paper/2512.14944/) (2025)

## Built on by (most cited first, 3 of 3)

- [AVIS: Adaptive Test-Time Scaling for Vision-Language Models](https://mlsanity.com/paper/2606.11576/) (2026)
- [PathoSage: Towards Multi-Source Evidence Adjudication in Pathology via Experience-Aware Agentic Workflow](https://mlsanity.com/paper/2606.07549/) (2026)
- [Sharpening Tax in Post-Training](https://mlsanity.com/paper/2610.01509/) (2026)

## Abstract (arXiv)

Reinforcement learning (RL) is increasingly used to post-train medical Vision-Language Models (VLMs), yet it remains unclear whether RL improves medical visual reasoning or mainly sharpens behaviors already induced by supervised fine-tuning (SFT). We present a controlled study that disentangles these effects along three axes: vision, SFT, and RL. Using MedMNIST as a multi-modality testbed, we probe visual perception by benchmarking VLM vision towers against vision-only baselines, quantify reasoning support and sampling efficiency via Accuracy@1 versus Pass@K, and evaluate when RL closes the support gap and how gains transfer across modalities. We find that RL is most effective when the model already has non-trivial support (high Pass@K): it primarily sharpens the output distribution, improving Acc@1 and sampling efficiency, while SFT expands support and makes RL effective. Based on these findings, we propose a boundary-aware recipe and instantiate it by RL post-training an OctoMed-initialized model on a small, balanced subset of PMC multiple-choice VQA, achieving strong average performance across six medical VQA benchmarks.